Altered metabolic function induced by Aβ-oligomers and PSEN1 mutations in iPSC-derived astrocytes

Abstract

Altered energy metabolism in Alzheimer's disease (AD) is a major pathological hallmark implicated in the early stages of the disease process. Astrocytes play a central role in brain homeostasis and are implicated in multiple neurodegenerative diseases. Although numerous studies have investigated global changes in brain metabolism, redox status, gene expression and epigenetic markers in AD, the intricate interplay between different metabolic processes, particularly in astrocytes, remains poorly understood. Numerous studies have implicated amyloid-β and the amyloid-β precursor in the development and progression of AD. To determine the effects of amyloid-β peptides or the impact of amyloid-β precursor protein processing on astrocyte metabolism, we differentiated astrocytes from induced pluripotent stem cells derived from people with early onset familial AD and controls. This study demonstrates that familial AD-derived astrocytes exhibit significantly more changes in their metabolism including glucose uptake, glutamate uptake and lactate release, with increases in oxidative and glycolytic metabolism compared to acute amyloid-β exposure. In addition to changes in major metabolic pathways including glutamate, purine and arginine metabolism and the citric acid cycle, we demonstrate evidence of gliosis in familial AD astrocytes highlighting a potential pathological hallmark. This suggests that chronic alterations in metabolism may occur very early in the disease process and present significant risk factors for disease progression for patients with early onset AD. These findings may also reveal important drivers of disease in late onset dementia and highlights key targets for potential diagnostic features and therapeutic agents in the future.

Keywords: Alzheimer's; astrocytes; gliosis; inflammation; metabolism; stem cells.

FIGURE 1

Characterisation of ‘healthy’ iPSC‐derived astrocytes. (a) Representative images showing differentiation and ICC staining of iPSC‐derived astrocytic cells at Day 45+. The astrocytes were differentiated from ‘healthy’ control NPCs for >40 days using astrocytes differentiation and maturation protocols. Top) Phase contrast images of control astrocytes on Days 2, 10 and 40 in culture post differentiation. Bottom) Cells were stained using immunofluorescent antibodies for astrocytic markers ALDH1L1 (left, green), S100β (middle, green) and GFAP (right, red), nuclei were counterstained in each image with DAPI (blue). Scale bars: 100 μM. (b) Cellular glycogen content of control astrocytes following exposure to hypoglycaemic conditions and treatment with DAB over 60 and 120 min. (c) Glycogen content of cells treated with dbcAMP an isoproterenol. (d) Glycogen contents of cells treated with Glutamate, oubain, glutamate and ouabain, TBOA and glutamate and TBOA for 60, 180 and 360 min. Results are expressed as ug/mg protein ± SD, n = 3 p < 0.05 (*), p < 0.01 (**), p < 0.001 (***). For DAB (b), a two‐way analysis of variance (ANOVA) was performed follow by Sidaks post‐test. Comparisons between treatments (c) were performed using analysis of variance (ANOVA) followed by Dunnet's post‐test. Each replicate or ‘n’ represents an independent culture preparation and is displayed as an individual data point.

FIGURE 2

Amyloid β‐treated control astrocytes exhibit difference in their bioenergetic profiles and metabolite processing. Astrocytes (45 days old) were treated with 0.2, 1 and 2 μM Aβ oligomers. (a) Profiles of Seahorse XFp Mito Stress Test data for oxidative consumption rates (OCR, pmol/min). (b) Percentage change of treated cells over control (untreated) cells. (c) Extracellular acidification rate (ECAR, mpH/min) measured after treatment with 0.2, 1, 2 μM Aβ oligomers. (d) Cellular glycogen content of cells (μg/mg cellular protein). (e) Glucose levels remaining in the media μM/mg cellular protein, (f) Glutamate uptake (nmol/mg cellular protein), (g) Lactate release (nmol/mg cellular protein) after treatment with 0.2, 1, 2 μM Aβ oligomers. Results are expressed as ± SD, n = 3 p < 0.05 (*), p < 0.01 (**), p < 0.001 (***). Comparisons between treatments were performed using ANOVA followed by Dunnet's post‐test. Each replicate or ‘n’ represents an independent culture preparation and is displayed as an individual data point.

FIGURE 3

Astrocytes display markers of gliosis following exposure to Aβ oligomers. Control astrocytes were treated with Aβ oligomers for 48 h. Astrocytes (45 days old) were treated with 0.2, 1, 2 μM Aβ oligomers. Cytokine or GFAP levels in media were measure using ELISA or by investigating intracellular accumulation using flow cytometry or ELISA. (a) IL‐6 levels (pg/mL) in the media were measured using ELISA or (c) Flow cytometry following treatment of astrocytes with Aβ oligomers. (b) IL‐8 levels (pg/mL) in the media were measured using ELISA or (d) Flow cytometry (fold change) following treatment of astrocytes with Aβ oligomers. (e) Secreted levels of GFAP in the media or (f) GFAP in cell lysates were measured using ELISA (ng/mL). Results are expressed as ± SD, n = 3 p < 0.05 (*), p < 0.01 (**), p < 0.001 (***). Comparisons between treatments were performed using ANOVA followed by Dunnet's post‐test. Each replicate or ‘n’ represents an independent culture preparation and is displayed as an individual data point.

FIGURE 4

Astrocytes process APP differentially in ‘healthy’ control versus fAD patient‐derived cells. Control and fAD (PSEN1 mutation)‐derived astrocytes (Day 45) were characterised using immunofluorescence and APP processing was assessed using ELISA and ADAM 10 activity. (a) Representative images of fAD (PSEN1 mutation) derived astrocytes. Cells were stained using immunofluorescent antibodies for astrocytic markers ALDH1L1 (green), S100β (Green) and GFAP (red). nuclei were counterstained with DAPI (blue). Scale bars: 100 μM. (b) Characterisation of ADAM10 enzymatic activity (RFU) and (c) soluble APPα (ng/mL) in control and PSEN1 (A246E, L268V and R278I). Pooled PSEN1 samples compared to control are displayed. (d) Aβ 1–40 (pg/mL), (e) Aβ 1–42 (pg/mL), (f) Aβ42/40 ratio, (g) aggregated Aβ (pg/mL) were measure in control and fAD patient‐derived astrocytes after 48 h. Pooled PSEN1 samples compared to control are displayed. Results are expressed as ± SD, n = 3 p < 0.05 (*), p < 0.01 (**), p < 0.001 (***). For direct comparison between control and PSEN1, unpaired t‐tests were performed. Comparisons between individual PSEN1 lines were performed using ANOVA followed by Dunnet's post‐test. Each replicate or ‘n’ represents an independent culture preparation and is displayed as an individual data point.

FIGURE 5

fAD‐derived astrocytes exhibit differences in their bioenergetic profiles and metabolite processing compared to ‘healthy’ control cells. (a) Profiles of Seahorse XFp Mito Stress Test data for oxidative consumption rates (OCR, pmol/min). (b) Extracellular acidification rate (ECAR, mpH/min). (c) Percentage change of fAD cells over control. (d) Glucose levels remaining in the media (μg/mg cellular protein), (e) Glutamate uptake (nmol/mg cellular protein) following addition of glutamate (200 μM) over 30, 60 and 120 min. Results are expressed as ± SD, n = 3 p < 0.05 (*), p < 0.01 (**), p < 0.001 (***). Comparisons between treatments were performed using ANOVA followed by Dunnet's post‐test. Each replicate or ‘n’ represents an independent culture preparation and is displayed as an individual data point.

FIGURE 6

fAD astrocytes display markers of gliosis compared to ‘healthy’ control astrocytes. Cytokine, GFAP, 8‐isoprostane levels in media or cellular lysates were measured in control and fAD astrocytes (45 days old). (a) Levels of GFAP were measure in the cell lysates or (b) cell culture media using ELISA (ng/mL). (c) Isoprostane levels were also measured in cellular lysates (pg/mg). Pooled control and fAD cell samples were compared in (d) cellular GFAP and (e) secreted GFAP, (f) isoprostanes. Fold change in the accumulation of (g) IL‐6 and (h) IL8 following protein transport inhibition measured using flow cytometry. Results are expressed as ± SD, n = 3 p < 0.05 (*), p < 0.01 (**), p < 0.001 (***). For direct comparison between control and PSEN1, unpaired t‐tests were performed. Comparisons between individual PSEN1 lines were performed using ANOVA followed by Dunnet's post‐test. Each replicate or ‘n’ represents an independent culture preparation and is displayed as an individual data point.

FIGURE 7

fAD‐derived astrocytes display significantly altered metabolic profiles compared to controls. (a) Heatmap displaying differential compounds identified using metabolomic analysis. Summary of pathway analysis for the comparison of control and b) A246E, (c) L286V and (d) R278 fAD patient‐derived astrocytes. The pathway analysis results of PSEN1 astrocytes compared with control. The colour graduated from white to and red indicates the degree of significance, the size of bubble represents the number of metabolites hit in the pathway. (e) Bar charts indicating intensity changes in key metabolites represented within metabolic pathway that were detected in astrocytes carrying PSEN1 mutations. Data shown are expressed as ± SD, n = 6. Each replicate or ‘n’ represents an independent culture preparation.